Today's mission:
Complete batch test of IRF540N power MOSFET in a power module and filter according to Vgs(th)<4V的器件。

Testing process:

1. Equipment construction

   • Using Keysight B2902A Precision Source Table

   • Custom test fixtures (Note: Probe oxidation is found to cause contact resistance deviation)

   • Ambient temperature 25℃±1℃ (the air conditioning system is unstable, and 0.8℃ fluctuates at 10:30)

2. Key test items

   • Threshold voltage Vgs(th): Vgs at Id=250μA

   • On-resistance Rds(on): Vgs=10V, Id=5A

   • Gate leakage current Igss: Vgs=±20V

Discovered problems:

• Sample 3# Vgs(th) drifts by 0.3V during repeated tests (initial value 3.8V → secondary test 4.1V)

• The Igss of sample 7# exceeds the standard (special book ≤100nA, actual measurement 230nA)

Fault analysis:

1. Electrostatic damage?

   • Check anti-static measures: the ion fan is normal, the wristband impedance is qualified

   • However, it was found that the operating table was not grounded with slight charge accumulation (static pen detection +12V)

2. Test method error

   • The pulse test method (100μs pulse width) and the DC test results are 8% different (the oscilloscope verifies that there is no oscillation)

   • Thermal effect causes the Rds(on) reading to rise slowly (infrared thermal imager shows that the shell and tube are ΔT=15℃ within 3 seconds)

Solution:

• Perform an X-ray check on the abnormal samples (discover the internal bond line offset of sample 3#)

• Use Tek DMM6500+3706 switch matrix to perform automated scanning (reduce human contact)

• Added pre-test aging link (the data tends to stabilize after 50 cycles of 5Vgs)

In-depth thinking:

1. The device specifications do not specify the correlation between test conditions and microscopic mechanism

   • For example, Vgs(th) actually corresponds to the critical point of the inverse layer formation, but the selection of Id threshold in the test directly affects the results

2. Differences in testing standards between commercial and industrial devices

   • This time, different batches of shelf samples were mixed, and it was found that the process variation coefficient of Batch B was 47% higher than that of 47%.

Tomorrow's plan:

• Design the temperature cycle test in the HALT test chamber (-40℃~85℃)

• Learn the power device testing specifications in the SEMI G32-88 standard

Technical Notes:
"The essence of testing is to build a bridge of dialogue between devices and standards. Today, it is realised again that there is a physical story behind every anomaly data—maybe the shout of lattice defects, or the slander of the oxide trap."

Features of this diary:

1. Includes engineering details such as specific device models, test parameters, etc.

2. Record fault phenomena → analyze → verification complete closed loop

3. Thinking at both technical details and methodological levels

4. Use professional terms but keep them readable

The specific content can be adjusted according to actual test items (such as ICs, sensors, passive components, etc.). What details need to be added can be told to me.